First post here.
At work yesterday undertaken a traverse to install lost control stations down the motorway with a Leica ms50 1?.?ÿ
Approx 18 stations were installed prior to starting, known backsite method was used from two existing control stations approximate loop length circa 2000m.
Everything seemed to be going well with no massive errors throughout the traverse but upon closing back to the two original stations the Hz dist was within 20-30mm and height 3mm ?ÿbut there was a great angular error amounting to almost 200+mm in position. Has anyone experienced anything like this before? All the usual things were checked, prism offsets,heights, co ordinates, atr check and adjust.
It could be the result of a linear traverse with angles close to 180?ø.?ÿ You should have some cross tie checks along the traverse.?ÿ If the traverse legs were close to 180?ø, I have experienced large closure errors like what you are seeing.?ÿ I would perform some GPS observations as a position check on a couple points at each end and the middle.
It's a weird one, unfortunately most on the control has been lost so checks with GPS will be the only one.
It could be the result of a linear traverse with angles close to 180?ø.?ÿ You should have some cross tie checks along the traverse.?ÿ If the traverse legs were close to 180?ø, I have experienced large closure errors like what you are seeing.?ÿ I would perform some GPS observations as a position check on a couple points at each end and the middle.
The only problem with angles close to 180 is that you can lose track of whether it was measured as an angle right or angle left, and that shows up as a serious closure error.
Nowadays everybody gets angle and distance of comparable accuracy, so it doesn't really matter what the angle is when using both values.
Of course, a short traverse leg will usually give a lot more angle error due to the effects of mis-centering, which will affect the angular closure a lot more than it does the positional closure.
Traverse geometry like that just itches for some cross ties. Say between the 417s and the 420s.?ÿ LS adjust the lot. And RTK vectors to a few of the stations, simultaneously adjusted with the traverse data, would not hurt a bit?ÿ?ÿ
Yes. You need cross ties.
I think the guys name was Hazelton who gave a presentation on how a control traverse almost exactly like the one shown created a huge headache and cost a lot of money to fix (and even more for the lawyers).
100% needs some cross ties, traverse route was decided by another engineer on site. Going to go back out and undertake something like below.
Only issue being majority of the stations are new?ÿ
Excellent analogy!. Jp
That will be a great improvement. But I would set on 417NB, BS 416NB (or 418NB) and tie 417 SB. Then I'd set on 417SB and tie 417 NB in similar fashion. Then repeat on the 420's. 4 set ups. The frame work you are proposing would take 8 setups, without really adding significantly more.
This reminds me of some control we inherited from another office a couple of years ago where the corridor was very narrow, zero cross-ties and adjusted by compass rule... we fixed it with some x-ties and Star*Net as I suspected a prism error somehow slipped into the original work. Star*Net is your friend in such circumstances.
I worked with a few folks who thought that every traverse and every level run had to be a literal loop on the ground, even if the width of the loop was a couple of feet (or even tenths). Sitting on the same stations going out and coming back was unthinkable and cross ties were pointless.
I just could not get them to see that half the number of stations with double the number of independent observations were better than single observations on a lot of stations. The loop closure was EVERYTHING...
...until someone input the wrong prism constant on a traverse similar to the OP's and held it as good because the closure was tight (equal number of stations on each side, ground distance was almost exactly the same, compensating error made it look great). After that mess was cleaned up, redundant measurements and cross ties started to get a bit more traction...
